Reports - CSAT

When to use it

• You want to track the satisfaction of the customers who talked to your operation.

• You want to understand how many CSAT surveys were sent and how many got a response.

• You want to compare satisfaction across different inboxes, teams, agents, or tags.

• You want to spot days with a drop in satisfaction and investigate why.

• You want to read the feedback each customer left in a conversation.

Prerequisites

• Be logged in as a supervisor, CX Engineer, or administrator.

• Have at least one conversation with a CSAT survey sent within the selected period.

Overview

The "Relatórios - CSAT" (Reports - CSAT) screen brings together, in a single place, all the information about the satisfaction surveys your operation sends at the end of conversations — including how many were answered, what the CSAT value was, and the written feedback the customer left.

The CSAT survey works in two moments:

1. Sending — at the end of a conversation, the survey is triggered to the contact.

2. Response — the customer picks a rating from 1 to 5 and, optionally, writes a comment.

Based on the ratings received, each response is classified into three groups:

These three groups are the basis for calculating the metrics shown on the screen.

What each block of the screen shows

The screen is divided into four main areas:

1. Filters at the top — let you slice the data by date, inbox, tag, team, and grouping period.

2. KPIs — Total responses, CSAT, and Response Rate.

3. Charts over time — CSAT evolution (with the total of positive, neutral, and negative responses) and response rate evolution.

4. Detailed table — one row per CSAT response received, with contact, agent, team, inbox, rating, tags, and feedback.

How the total cards are calculated

The three cards at the top of the screen answer three different questions.

Total de Respostas (Total Responses)

How many CSAT surveys were answered in the period. Each response corresponds to a conversation in which the customer picked a rating from 1 to 5.

CSAT

The percentage of positive responses (ratings 4 and 5) out of the total responses received:

CSAT (%) = Positive responses ÷ Total responses × 100

The higher it is, the greater the share of customers leaving satisfied. Rating 3 (neutral) and ratings 1–2 (negative) do not count as satisfaction in the CSAT calculation — they only pull the number down by increasing the denominator.

Taxa de Resposta (Response Rate)

The percentage of surveys that were answered out of the total surveys that were sent to the customer:

Response Rate (%) = Answered surveys ÷ Sent surveys × 100

This number shows how engaged with the survey your base is.

Difference between "survey sent" and "survey answered"

• Sent: at the end of the conversation, the system triggered the survey to the contact.

• Answered: the contact opened the survey, picked a rating, and submitted it.

Every answered survey was, before that, a sent survey — but the opposite is not true. It is common for sent surveys to get no response.

CSAT over time

The "CSAT ao longo do tempo" (CSAT over time) chart shows, for each period (day, week, month), two pieces of information overlaid:

• Stacked bars with the number of positive (green), neutral (yellow), and negative (red) responses.

• Line with the period's CSAT (%), calculated as positive responses ÷ total responses × 100.

How to read it

• The total size of the bar shows how many responses the period received.

• The proportion of the colors shows the distribution between positive, neutral, and negative.

• The line shows the evolution of CSAT (%) over time — useful for spotting trends of improvement or decline.

Response rate over time

The "Taxa de resposta ao longo do tempo" (Response rate over time) chart shows, for each period:

• Bars with the total responses received in the period.

• Line with the response rate (%) — answered ÷ sent × 100.

Detailed table

The table at the bottom of the screen brings one row per CSAT response. Each column helps put the rating in context:

The difference between Data da Resposta (Response date) and Data da criação do ticket (Ticket creation date) shows how long the customer took to answer the survey after the conversation started. It is common for responses to take a few hours (or even days) to arrive — that is why the CSAT of the most recent days may keep changing over the following days.

Available filters

The screen lets you cross responses by:

• Data de Criação do Ticket (Ticket creation date) — the period in which the conversation was created (not the date the survey was answered). It accepts predefined ranges (last 7 days, 30 days, etc.) and custom ranges.

• Caixa (Inbox) — inbox (channel) where the conversation took place.

• Tag — tag applied to the conversation.

• Time (Team) — team (currently) responsible for the conversation.

• Agrupar por período (Group by period) — defines the granularity of the evolution charts (Day, Week, Month).

Practical example

Scenario

You are looking at the screen with the "last 7 days" filter for a specific inbox.

Result

How to interpret it

• 7,195 responses for the conversations created in the last 7 days.

• CSAT of 71.73% → about 7 out of every 10 customers who answered left satisfied. The other 3 were either neutral or dissatisfied.

• Response rate of 19.02% → about 1 out of every 5 surveys sent was answered. The remaining 81% simply did not answer.

To understand what is pulling CSAT down:

1. Look at the chart over time and identify the day(s) with the most red bars.

2. Add filters to assess whether the dissatisfaction is concentrated in a specific topic.

3. Open the conversations listed to read the written feedback and understand what happened.

Frequently asked questions

Why is my response rate so low?

It is not unusual to have low response rates on CSAT surveys. Most customers simply do not answer — it does not mean the service was bad. If your rate is far below what you expected, it is worth reviewing:

• The channel in which the survey is sent.

• The message that goes with the survey.

Can I trust the CSAT when the response rate is very low?

With care. If only 5 customers answered in the period, any low rating weighs too much on the result. Use the number of responses as a reference: the higher the volume, the more reliable the period's CSAT is.

Can the numbers on this screen change over time?

Yes. A conversation closed today may have its survey answered tomorrow or later. Since the screen uses the ticket creation date as the reference, the metrics for the most recent days keep growing in the following days, as customers answer.

Looking at the current day's CSAT right after conversations are closed usually underestimates both the volume of responses and the consolidated CSAT. Wait a few days before treating the numbers as final.

Why do some rows in the table show up with the "no_tag" tag?

It means the corresponding conversation did not receive any tag during or after the service. It is worth reviewing with the team whether tagging is being done consistently.

Does the "Data de Criação do Ticket" (Ticket creation date) filter consider when the conversation was opened or when the survey was answered?

It considers when the conversation was created. This means that, when you filter by the "last 7 days", you are looking at surveys from conversations started within those 7 days — some of them may have been answered days later.

Do answered surveys without written feedback show up too?

Yes. The written feedback is optional. Every rating given counts as a response, even without a comment. In the detailed table, the feedback column stays empty in these cases.

Why is the total number of responses different from one chart to another?

At times, the total responses may show up slightly different between the charts. This happens temporarily and tends to sort itself out over the course of the day.

There are two main reasons:

• Updates at different moments — the survey information and the conversation information are updated every hour, but not at exactly the same instant. If you open the screen right in the window where one has already been updated and the other has not, the numbers may not match 100%.

• Cached view — to make the screen faster, part of the content is stored for a few minutes. It may happen that one chart shows the most recent version and another still shows the previous one.

In both cases, the difference is usually small and resolves itself on the next update.

Notes

• To understand other metrics of your operation, check the other articles in the Metrics and reports category.